A blood-based mRNA signature distinguishes people with Long COVID from recovered individuals, 2024, Missailidis, Annesley+

[SNT Gatchaman]

A blood-based mRNA signature distinguishes people with Long COVID from recovered individuals
Missailidis, Daniel; Ebrahimie, Esmaeil; Mohammadi-Dehcheshmeh, Manijeh; Allan, Claire; Sanislav, Oana; Fisher, Paul Robert; Gras, Stephanie; Annesley, Sarah Jane

Long COVID is a debilitating condition that lasts for more than three months post-infection by SARS-CoV-2. On average, one in ten individuals infected with SARS-CoV-2 develops Long COVID worldwide. A knowledge gap exists in our understanding of the mechanisms, genetic risk factors, and biomarkers that could be associated with Long COVID.

In this pilot study we used RNA-Seq to quantify the transcriptomes of peripheral blood mononuclear cells isolated from COVID-recovered individuals, seven with and seven without Long COVID symptoms (age-and sex-matched individuals), on average 6 months after infection.

Seventy genes were identified as significantly dysregulated up-or down-regulated in Long COVID samples, and the vast majority were downregulated. The most significantly dysregulated up or downregulated genes fell into two main categories, either associated with cell survival or with inflammation.

This included genes such as ICOS (FDR p = 0.024) and S1PR1 (FDR p = 0.019) that were both upregulated, indicating that a pro-inflammatory state is sustained in Long COVID PBMCs compared with Post-COVID recovered PBMCs. Functional enrichment analysis identified that immune-related functions were expectedly predominant among the dysregulated up-or down-regulated genes.

The most frequently downregulated genes in significantly altered functional categories were two leukocyte immunoglobulin-like receptors LILRB1 (FDR p = 0.005) and LILRB2 (FDR p = 0.027). Downregulation of these inhibitory receptors similarly indicates a sustained proinflammatory state in Long COVID PBMCs.

PCA analysis demonstrated that LILRB1 and LILRB2 expression discriminated all of the Long COVID samples from Post-COVID recovered samples. LILRB1 and LILRB2 should be validated as prospective biomarkers of Long COVID in larger cohorts, over time and against clinically overlapping conditions.

Link | PDF (Frontiers in Immunology)

 

[Sean]

The most significantly dysregulated up or downregulated genes fell into two main categories, either associated with cell survival or with inflammation.

Interesting.

 

[Hutan]

PCA analysis demonstrated that LILRB1 and LILRB2 expression discriminated all of the Long COVID samples from Post-COVID recovered samples. LILRB1 and LILRB2 should be validated as prospective biomarkers of Long COVID in larger cohorts, over time and against clinically overlapping conditions.

LILRB1 - also known as ILT2; LIR1; MIR7; PIRB; CD85J; ILT-2; LIR-1; MIR-7; PIR-B

The receptor is expressed on immune cells where it binds to MHC class I molecules on antigen-presenting cells and transduces a negative signal that inhibits stimulation of an immune response. It is thought to control inflammatory responses and cytotoxicity to help focus the immune response and limit autoreactivity. Multiple transcript variants encoding different isoforms have been found for this gene.​

 

[DMissa]

I have noticed that the abstract is cooked (words mashed together several times) no idea whether this came from the journal or from changes of mine during proof stage (possibly while I was hanging by a thread between home visits) but I have written to them to rectify if possible.

I am a pretty fastidious person so I thought this was worth acknowledging. The last thing I want is for last minute abstract changes to unduly convey a lack of care in the actual work!

This is a small study where we used seed funding to investigate some initial samples from the COVID biobank run by Prof Stephanie Gras at LTU. Expansion + lots of future work and collaborations planned or underway. Really exciting breadth of experts involved across several projects big and small. This is barely a tip of the iceberg. It is great that top tier researchers are taking the field so seriously.

 

[Kitty]

Expansion + lots of future work and collaborations planned or underway.

Exciting!

It's good to see a team being really inventive and squeezing maximum value from small pots of money.

 

[DMissa]

Exciting!

It's good to see a team being really inventive and squeezing maximum value from small pots of money.

Very exciting. I should also clarify that a lot of the work happening is not just that driven by us or our study here. So many researchers are bringing their own new ideas into this space and designing incredibly innovative studies. People are genuinely interested in this and it’s a real breath of fresh air. Some massive team collaborations happening.

It will be good for ME/CFS research as well, I think (not to necessarily treat these as separate issues, there will be overlaps and differences both, as we know). The Mason-funded ME/CFS study that I am currently doing home visits for has brought LC researchers into ME/CFS, for example. I am really hopeful for the future of the field overall.

ME/CFS vs LC cell biology comparison work is some of what I am most excited about, too. I presented some at Oxford in April, hopefully it matures and gets into publication next year.

 

[EndME]

Thanks for working on this and in particular thank you for commenting here and by doing so engaging with patients and researchers @DMissa.

Thanks for providing some cohort data. Unfortunately, I didn't see anything on comorbidties, was this not examined? Do you know if the cohorts were BMI matched or if there were big differences in BMI? I noticed that 60% of patients are reporting a cough, which seems very unusual to me, do you have any reasonable explanation for this or is this a rather unusual cohort characteristic (perhaps being to close to the initial infection or a cohort that suffered more from lung damage or a cohort that has certain comorbidities).

I must admit that over the years, "prospective biomarkers for Long-Covid" have very much become a meme for me, having by now probably read hundreds of papers that have claimed to have identified "prospective biomarkers" that never headed anywhere. Most of which were very likely just noise caused by cohort selections and measurement/methodological noise.

If you can excuse me for being a bit provocative (I'm not trying to be discouraging, rather the opposite): Why do you think this work could be different, respectively why have others not found what you have found here (especially in relation to studies that have had a more robust sample size)?

 

[CorAnd]

Seventy genes were identified as significantly dysregulated up-or down-regulated in Long COVID samples, and the vast majority were downregulated. The most significantly dysregulated up or downregulated genes fell into two main categories, either associated with cell survival or with inflammation.

It would be so interesting to know if the same genes are downregulated in ME/CFS. Will maybe DecodeME answer this question?

 

[DMissa]

Thanks for providing some cohort data. Unfortunately, I didn't see anything on comorbidties, was this not examined? Do you know if the cohorts were BMI matched or if there were big differences in BMI? I noticed that 60% of patients are reporting a cough, which seems very unusual to me, do you have any reasonable explanation for this or is this a rather unusual cohort characteristic (perhaps being to close to the initial infection or a cohort that suffered more from lung damage or a cohort that has certain comorbidities).

Cohort is too small to subdivide and get much meaningful out of it, I think. Certainly, stats aren't going to work. Important for future expansion, though, yes.

I must admit that over the years, "prospective biomarkers for Long-Covid" have very much become a meme for me, having by now probably read hundreds of papers that have claimed to have identified "prospective biomarkers" that never headed anywhere. Most of which were very likely just noise caused by cohort selections and measurement/methodological noise.

If I instead wrote "possible discriminator" I would sidestep the meme but this would not change anything about the work. As for other research not getting to the end of the road yet, I would not make assumptions about what everybody has cooking behind closed doors or the reasons for things not yet eventuating into publication.

If you can excuse me for being a bit provocative (I'm not trying to be discouraging, rather the opposite):

You don't need to explain, I understand. Accountability is good and there are obvious limitations to this very small piece of work.

Why do you think this work could be different, respectively why have others not found what you have found here (especially in relation to studies that have had a more robust sample size)?

I do not think that similar work has hit a dead end. I think that it is still early days. Research moves much more slowly than the few years we have had since the pandemic. As for why LILRB1 and LILRB2 haven't stood out as sharply in larger studies, the best reasons that aren't "our cohort is small and there could be skewed representation" are: there could be differences in case ascertainment and there are differences in the bioinformatics approaches. I think unlikely that is is random noise, looking at the scatter plots and considering how functionally related both genes are. Seems striking to me.

Anyway these are really all of the obvious things. I am sure that you know them. There's no magic explanation. I'm not going to handwave. It's a small study undertaken with a tiny amount of seed funding. We have said as much. The important thing is, that, in my opinion, when you take into account function, the key results are interesting enough to be worth testing further. Validation can be done cheaply and easily which is the good thing here. Isolate some PBMCs and do some PCRs. It's really not going to lead us down a rabbit hole of wasted time and funding. Do I think that it is still going to be 100% accurate in 1000 people? The answer is obvious, probably not. Do I think there might be something useful here to understand in more detail? Yes. Are there mechanistic angles to pursue here (that do match where the rest of the field seems to be heading)? Yes.

And, again, we have a lot more cooking, this is a crumb of our ongoing and planned ME/CFS - LC research. This is our lab getting our toe in the LC publication door in the interim. I am very proud of what some pretty special people in the team have extracted from the small amount of resourcing that was provided for this study when it was originally done. We are becoming part of much larger things now, and this was a big help towards getting there.

It would be so interesting to know if the same genes are downregulated in ME/CFS. Will maybe DecodeME answer this question?

We are following this (and other experiments) up with bigger cohorts from both groups (I say "both" but there may obviously be overlap).

 

[Hutan]

Thanks for working on this and in particular thank you for commenting here and by doing so engaging with patients and researchers @DMissa.

Yes, thank you from me too.

6/7 participants were female in the LC cohort; only 5/7 participants were female in the control cohort.

The most reported symptoms were fatigue, post-exertional malaise (PEM), muscle pain, concentration and memory issues, sore throat, headaches and temperature dysregulation. Fatigue was reported by each of the Long COVID participants and five out of the seven reported PEM. Fatigue and PEM were rated with the greatest overall severity of all symptoms

great to see PEM being assessed in an LC study.

Total RNA was isolated from PBMCs of all participants and sent for RNA-Seq analysis

PBMC transcriptome
Possibly different percentages of PBMCs could account for the different expression of LILRB1 and LILRB2.

As for why LILRB1 and LILRB2 haven't stood out as sharply in larger studies, the best reasons that aren't "our cohort is small and there could be skewed representation" are: there could be differences in case ascertainment and there are differences in the bioinformatics approaches. I think unlikely that is is random noise, looking at the scatter plots and considering how functionally related both genes are. Seems striking to me.

It certainly is interesting that the two related genes stood out when it comes to expression.
Is it possible that there was effectively some double-counting, that is, the bits of the proteins you were quantifying mean that there is some inaccuracy in identifying the two different proteins?

I'm not quite following Figure 2b, the heat map of differential expression for the two cohorts.

(B) Heatmap representation of seventy differentially expressed genes between Long COVID and COVID recovered PBMC samples. Red indicates higher expression while blue indicates lower expression. Each column is a different sample and each row is a different gene.

The scale goes from 0 (blue) through to 50 (red). What is the comparison point? I thought it might be the average expression for the control cohort. But then, wouldn't you have to have a scale that goes from negative numbers to positive numbers? Are the white squares similar to the yellow squares or do they indicate missing data?
I'm puzzled, for example, how CD14 can be a down regulated gene, but almost all of the Long Covid participants have red (high) levels of gene expression, while most of the control participants have white levels, and two have lower levels of gene expression than most of the LC participants.

 

[Hutan]

All immunogloblin genes are located in chromosome fourteen, and as a group constituted the most frequently downregulated class of proteins in the Long COVID PBMCs. Seven out of the sixty-six downregulated genes belong to this class, namely IGLV8-61, IGHV1-18, IGHV3-48, IGHV2-5, IGHV4-59, IGHV3-7, and IGHV3-30. It is important to note that IGLV8-61 is a secretory protein and is predicted to be active in the extracellular space.

We've seen immunoglobulin genes mentioned in other studies. Possibly though there are so many that it's not surprising that some look different?

There was a recent Scheibenbogen study that reported this:

IG - immunoglobulin proteins
increased - IGHV1-2
decreased - IGKVA-1, IGKV1-17, IGHV5-51, IGLV8-61, IGHV3-7

Only IGLV8-61 is in common, but at least both studies found its expression was decreased.

 

[Hutan]

An analysis of transcriptomic data from nine different studies that compared SARS-CoV-2 infected cohorts to either healthy controls or other respiratory diseases identified that the expression of IGHV genes was most consistently up-regulated specifically in severe SARS-CoV-2 infections (33). This is noteworthy since severe infections are the most associated with development of Long COVID. A shift from elevated IGHV expression during the acute phase to reduced IGHV expression in Long COVID suggests dysregulation, and perhaps overactivity of the regulatory pathways involved in bringing the immune response back down over time.

I think the bolded sentence is only true if the definition of Long Covid is very broad. If Long Covid is, as it should be, stratified to separate out people with post-viral fatigue-type illness from those with various other consequences of a Covid-19 infection, then I don't think there is good evidence that the post-viral fatigue-type illness is more common in those with severe infections.

We need Long Covid researchers to give a lot of attention to trying to get fairly homogeneous cohorts. Finding a common mechanism for lung damage resulting from severe Covid-19 (including as a result of the treatment applied) and ME/CFS-like LC is probably not possible.

 

[Hutan]

Reduced expression or impaired function of LILRB1 and LILRB2 have been associated with inflammatory autoimmune conditions such as rheumatoid or psoriatic arthritis and systemic lupus erythematosus (49–52).

As well as the 4 references cited there, there is this one:
A LILRB1 variant with a decreased ability to phosphorylate SHP-1 leads to autoimmune diseases
It found that a family with incidences of Graves’ disease, Hashimoto's thyroiditis, systemic lupus erythematosus had a variation of LILRB1 that decreased its function.

 

[DMissa]

PBMC transcriptome
Possibly different percentages of PBMCs could account for the different expression of LILRB1 and LILRB2.

I think there is immunotyping coming up in the larger studies which would be good

Is it possible that there was effectively some double-counting, that is, the bits of the proteins you were quantifying mean that there is some inaccuracy in identifying the two different proteins?

They are close together on the chromosome so it is likely that a factor regulating expression of genes in this region is doing something abnormal that is affecting both of them.

Re: heatmap questions, requires me to look in more detail to answer with 100% certainty, and possibly consult co-authors who did the bioinformatics + figure generation, perhaps legend could have been more detailed. Don't have the time today but I'll try to get to it. Off to pick up a blood sample